TL;DR
What did monkeydust use ElevenLabs for?
Not wanting their own voice on a video, a creator ran the recording through ElevenLabs speech-to-speech instead of text-to-speech, which kept their natural pauses and broken sentences in the new voice. They showed it to a group of people who "could not believe it was non-human." They say the pauses and broken sentences, which a typed script usually leaves out, were what made it convincing.
What they did
They recorded the narration themselves, uploaded the video to ElevenLabs speech-to-speech, picked a different voice and adjusted a few settings. It replaced the audio track with the new voice while keeping their delivery.
How it went
They showed it to a group of people who "could not believe it was non-human." They say the pauses and broken sentences, which a typed script usually leaves out, were what made it convincing.
Worth knowing
- Use a voice you have the rights to. Some platforms require you to label AI-altered voices.
- Price check: ElevenLabs: free plan; Starter: US$6/month, or US$5/month on a yearly plan (US$60/year) (checked Nov 1, 2026).
Try it yourself
Paste this into a chatbot, then add your own details underneath.
Record your narration naturally, with your usual pauses. In ElevenLabs speech-to-speech, upload the clip, pick a voice from the library, keep the settings close to default, and compare a short section before converting the whole video.
Original post
See monkeydust's original post on Hacker News → (Jul 8, 2026)
Sources
- monkeydust on Hacker News, Jul 8, 2026 — the original post
- ElevenLabs pricing — checked Nov 1, 2026